Controlling Au Photodeposition on Large ZnO ... - ACS Publications

May 19, 2016 - Centre for Advanced Imaging, University of Queensland, Brisbane, Queensland ... Band-gap photoexcitation of electronic charge in ZnO by...
0 downloads 0 Views 4MB Size
Download PDF
Research Article www.acsami.org

Controlling Au Photodeposition on Large ZnO Nanoparticles Joseph F. S. Fernando,† Matthew P. Shortell,† Christopher J. Noble,‡ Jeffrey R. Harmer,‡ Esa A. Jaatinen,† and Eric R. Waclawik*,† †

School of Chemistry, Physics and Mechanical Engineering, Science and Engineering Faculty, Queensland University of Technology, Brisbane, Queensland 4000, Australia ‡ Centre for Advanced Imaging, University of Queensland, Brisbane, Queensland 4072, Australia S Supporting Information *

ABSTRACT: This study investigated how to control the rate of photoreduction of metastable AuCl2− at the solid−solution interface of large ZnO nanoparticles (NPs) (50−100 nm size). Band-gap photoexcitation of electronic charge in ZnO by 370 nm UV light yielded Au NP deposition and the formation of ZnO−Au NP hybrids. Au NP growth was observed to be nonepitaxial, and the patterns of Au photodeposition onto ZnO NPs observed by high-resolution transmission electron microscopy were consistent with reduction of AuCl2− at ZnO facet edges and corner sites. Au NP photodeposition was effective in the presence of labile oleylamine ligands attached to the ZnO surface; however, when a strong-binding dodecanethiol ligand coated the surface, photodeposition was quenched. Rates of interfacial electron transfer at the ZnO− solution interface were adjusted by changing the solvent, and these rates were observed to strongly depend on the solvent’s permittivity (ε) and viscosity. From measurements of electron transfer from ZnO to the organic dye toluidine blue at the ZnO−solution interface, it was confirmed that low ε solvent mixtures (ε ≈ 9.5) possessed markedly higher rates of photocatalytic interfacial electron transfer (∼3.2 × 104 electrons·particle−1·s−1) compared to solvent mixtures with high ε (ε = 29.9, ∼1.9 × 104 electrons·particle−1·s−1). Dissolved oxygen content in the solvent and the exposure time of ZnO to band-gap, near-UV photoexcitation were also identified as factors that strongly affected Au photodeposition behavior. Production of Au clusters was favored under conditions that caused electron accumulation in the ZnO−Au NP hybrid. Under conditions where electron discharge was rapid (such as in low ε solvents), AuCl2− precursor ions photoreduced at ZnO surfaces in less than 5 s, leading to deposition of several small, isolated ∼6 nm Au NP on the ZnO host instead. KEYWORDS: ZnO, Au, Au−ZnO nanoparticle hybrids, nanoparticle composites, photodeposition, photochemical reduction, photocatalysis, interfacial electron transfer

1. INTRODUCTION The attachment of noble metal nanoparticles (NPs) to semiconductor NPs is a powerful approach to producing new or modified optical phenomena1,2 in advanced materials that can also gain new chemical functionality thereby.3 This form of materials combination also generates enhanced photoelectrochemical activity.4 For example, when Au NPs are attached to a UV-active semiconductor metal oxide (MOx) such as TiO2 or ZnO, the new hybrid materials’ light absorption wavelength ranges are extended, and photoinduced charge transfer is enhanced.5 These phenomena have frequently been examined in studies of photodegradation of chemical pollutants and photocatalytic fine chemical production by hybrid NPs.6−8 One reason why photocatalytic charge transfer can be efficient in MOx semiconductor−metal hybrids is because metal NP © XXXX American Chemical Society

loading can circumvent the barrier to charge transfer caused by the space-charge layer that naturally forms when an n-type semiconductor like TiO2 or ZnO is placed in contact with an electrolyte.9 The formation of a space-charge layer at the semiconductor−solvent interface normally suppresses rates of photoexcited electron transfer to species adsorbed onto the MOx NP surface.4,8 Another difference of these hybrids is that recombination of photoexcited carriers is reduced, a process that competes directly with photochemical reduction.9 Attachment of noble metal NPs to a MOx semiconductor can increase the catalytic activity by promoting charge separation of excitons Received: March 14, 2016 Accepted: May 19, 2016

A

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

efficiently transfer and are partitioned into the attached Au, where they can accumulate and participate in photocatalysis.5 This investigation focused on achieving control over photodeposition of Au NPs onto large (80 nm) prolate ZnO NPs, principally by changing interfacial electron-transfer rates through changes to external conditions. Due to their large size, these ZnO NPs possessed some traits similar to bulk ZnO and ZnO semiconductor thin films. ZnO is a photochemically active, n-type semiconductor material with a wide, direct band gap of Eg(298 K) = 3.365 ± 0.005 eV, which corresponds to a wavelength of approximately 365 nm.35 In its macroscopic form and at room temperature (298 K), ZnO possesses an exponential absorbance tail to the low-energy side of the first exciton resonance, the Urbach tail.36 Thus, at 298 K, ZnO absorbs light up to an optical cutoff in the near-UV at approximately 380 nm. 35,37 ZnO also has rich defect chemistry.38 Intrinsic lattice defects such as Zn vacancies, interstitials, and oxygen vacancies naturally arise during ZnO synthesis, and when present in high concentration, these defects generate sub-band-gap states that can be examined by techniques such as electron paramagnetic resonance (EPR).39−41 and visible fluorescence emission.42 Defects in ZnO semiconductor NPs are localized, high-energy surface traps for electrons and thus sites where photodeposition growth of the metal component can occur.32 It is logical that a metal precursor is more likely to be reduced where high-energy electrons accumulate on the semiconductor surface. Studies of metal NP photodeposition on anisotropic ZnO and CdS that terminate in sharp tips have found that tip-growth of the metal occurs on these NPs and nanorod structures.32,43,44 The nanostructure tips of these noncentrosymmetric, wurtzite-phase materials are likely to have low coordination number, a high number of dangling bonds, and increased electron density compared to other sites on these nanostructures.44 The localization of electrons at the ends of ZnO nanorods prior to the electron-transfer step has been demonstrated to produce site-specific photodeposition of silver NPs.27 Preferential photodeposition of Pt at the defect sites on CdS nanorods has been demonstrated to occur in solution at room temperature,45 so tip- or edge-selective growth on faceted NPs might be expected in ZnO NPs also. Understanding where photogenerated electrons are distributed in the host NP requires knowledge of defect states and their locations. Controlling charge accumulation and interfacial charge transfer at these sites might be a way to control the growth locations of the metal component on the semiconductor NP surface. Although electron storage in MOx semiconductor NPs, ZnO in particular, has previously been investigated,46 these recent studies have focused on photodoping and the IR and EPR signature of delocalized conduction band electrons produced in the process.47 The majority of studies of charge storage in ZnO appear to have focused on ZnO QDs and small NPs.28,29,46−48 Recently, Gamelin and co-workers47 confirmed that, in small ZnO NPs, the maximum number of electrons that can be stored per NP is particle-size-dependent. They also demonstrated that the maximum number of electrons stored on ZnO NPs strongly depends on the nature of the hole quencher used in the colloid system, because the kinetics of ZnO h+VB quenching determines the maximum electron density supported by ZnO. For a given hole quencher, the maximum ZnO e−CB storage density is also essentially constant over a large range of ZnO NP sizes. For this reason, the large-sized ZnO NPs examined in this study were expected to store greater

(bound electron−hole pairs), generated in the semiconductor following band-gap light absorption, producing free, highenergy charge carriers.10,11 If it is assumed that charge transfer from the semiconductor to the attached metal is efficient, the different components of the hybrid can accumulate the separated charges, which in the case of electrons then possess sufficient reduction potential to reduce certain functional groups on a given target chemical substrate.5 Plasmonic effects in attached noble metal NPs can also be exploited in these photocatalyst systems. When semiconductor-to-metal charge transfer is not efficient, direct light absorption by the noble metal and excitation of the metal’s free electrons to higher energy states can perform this photocatalytic function.6,8 There are a number of ways by which Au or other noble metal NPs can be loaded onto MOx and other semiconductor nanostructures.3,12−14 For instance, the two components of a hybrid can be prepared separately and then linked together with a suitable ligand, such as a bidentate thiol linker or similar molecule.11,15,16 This usually leads to an insulating organic layer between the metal and semiconductor that hinders charge transport.13,15 Au NPs can be used as crystallization seeds that, upon addition to a supersaturated solution of a precursor salt solution of the second component, can generate a wealth of interesting hybrid structures.2,12 Photodeposition of the metal onto the semiconductor nanostructure is another way that metal loading can be achieved, where light is absorbed by the semiconductor to generate high-energy electrons and reduce a metal salt at the semiconductor−solution interface.17−27 Each of these different synthesis approaches offers different levels of control over the number of components that combine to form the new hybrid NP, NP composition, shape, and interfacial connection. Photoexcitation and its effect on the physical properties of metal−semiconductor hybrids have been examined previously;28−31 however, these early studies of Au, Ag, Cu, and Pt metal island growth on ZnO28 and TiO229−31 did not examine the question of how to define the location of metal NP deposition on the semiconductor because they were performed on small NP structures known as quantum dots (QD). Control of the position of metal NP photodeposition on the semiconductor and possible influence of semiconductor shape (and lattice matching), as well as metal NP size and shape, are all topics of interest when host particles of an intermediate size are used.27,32 The metal photodeposition route could prove advantageous where precise control over the metal domain size is required, which in the case of noble metal NPs may also prove to be a way to fine-tune plasmonic properties.32,33 Once the noble metal has been deposited at the surface of a semiconductor to form a hybrid NP, it can serve as an electron sink. UV irradiation of the hybrid NP generates electron−hole pairs that separate at the metal−semiconductor interface with the electrons accumulating in the metal, causing further metal reduction and growth at the deposition site.34 For the particular example of Au−ZnO hybrids, charge transfer from the semiconductor to the noble metal might normally be expected to be limited due to formation of a Schottky barrier at the interface, since the work function of Au (5.1 eV) is higher than the conduction band edge of ZnO (∼4.25 eV). However, recent Kelvin probe force microscopy measurements have shown that an Ohmic contact forms at the ZnO−Au interface in the dark. Under conditions of steady-state UV illumination, photogenerated electrons in the ZnO NPs’ conduction band B

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

obtained products were separated by centrifugation, washed several times with methanol, and dried in air. Dodecanethiol (DDT)-capped ZnO was prepared by passivation of uncapped ZnO. Typically, 20 mg of uncapped ZnO was dispersed in a 1:1 (by volume) mixture of ethanol and toluene containing 3 mmol of DDT, and the reaction mixture was stirred at room temperature overnight. Then the particles were separated by centrifugation, before being washed several times with ethanol and dried in air. 2.3. Photodeposition of Au Nanoparticles. All photodeposition reactions were carried out by use of a miniature UV diode (370 ± 10 nm), which illuminated the entire sample (four-sided 1 cm glass cuvette). The power and intensity incident on the cuvette were approximately 64 mW and 35 mW/cm2, respectively. Colloidal suspensions of ZnO NPs (0.4 mg/mL) were prepared in the desired solvent (ethanol, toluene, acetonitrile, butanol, t-butanol, cyclohexane, methanol, ethyl acetate, or water). HAuCl4 (0.1375 mM) precursor solutions were always prepared in ethanol. Ethanolic HAuCl4 is stable for several weeks in the dark. A typical synthesis involves three steps: (1) preirradiation of HAuCl4 solution for 6 min, (2) addition of 0.6 mL of preirradiated HAuCl4 to 1.25 mL of ZnO colloid in a glass cuvette, and (3) irradiation of the mixture for 20 s. Two irradiation techniques were used: “pulsed” (4 × 5 s exposure to 370 nm irradiation with a 30 s gap between pulses) and “continuous” (a single 20 s exposure to 370 nm irradiation). The diode was manually operated with a toggle switch. In order to evaluate the effect of dissolved oxygen in the photodeposition process, the two reactants were bubbled for 20 min with Ar gas or O2 gas before mixing and UV irradiation. 2.4. Photocatalytic Reduction of Toluidine Blue. As with Au photodeposition, ZnO NPs (0.4 mg/mL) were dispersed in the desired solvent mixture. Toluidine blue (TB+, 0.1375 mM) was prepared in ethanol. Thereafter TB+ and ZnO colloid were mixed, and the mixture was irradiated with 370 nm photodiode output for up to 1 min while the TB+ concentration was monitored spectroscopically. 2.5. Particle Characterization. Structure, morphology, and crystalline nature of particles were characterized by (high resolution) transmission electron microscopy [(HR)TEM], ultraviolet−visible extinction spectroscopy (UV−vis), electron paramagnetic resonance (EPR) spectroscopy and dynamic light scattering (DLS). HRTEM imaging was performed on a JEOL JEM-2100 instrument with an acceleration voltage of 200 kV, while low-resolution TEM images were recorded on a JEOL JEM-1400 instrument with an acceleration voltage of 100 kV. Samples for TEM and HRTEM were prepared by dropcasting a dilute solution of NPs dispersed in ethanol onto a carboncoated copper grid. In situ UV−vis spectra were obtained by use of an Ocean Optics HR4000CG-UV-NIR spectrometer. X-band (ca. 9.77 GHz) continuous-wave (CW) EPR spectra were recorded on a Bruker Biospin Elexsys E500 EPR spectrometer fitted with a super-high-Q cavity (CW EPR). Spectra were recorded under the following conditions: room temperature, modulation frequency 100 kHz, modulation amplitude 0.1 mT, and microwave power 20 mW (10 dB). Magnetic field and microwave frequency were calibrated with a Bruker ER 036TM Teslameter and a Bruker microwave frequency counter, respectively. DLS measurements for ZnO colloidal samples dispersed in different solvents were performed with a Malvern Zetasizer Nano S system.

amounts of electronic charge per particle than ZnO QDs. Dissolved oxygen in the solvent efficiently scavenges electrons, and so it also effects the kinetics of photochemical electron transfer during the deposition process. Since stored electrons on large ZnO NPs react only rather slowly with oxygen, with the rate constant becoming slower with increasing particle size, photodeposition on large ZnO NPs was expected to be more efficient compared to that on small ZnO NPs or ZnO QDs in ambient conditions.46 Under UV excitation, electrons generated in the ZnO conduction band were also expected to populate defect sites with sufficiently high reduction potential to convert Au3+ to Au0, leading to Au NP growth at the ZnO surface.44 The static relative permittivity (ε) of the solvent matrix can be expected to influence interfacial transfer of accumulated charge during photodeposition and photocatalytic processes,49 but this does not appear to have been extensively studied. Controlling photodeposition of metal NPs onto semiconductor particles by tuning photoexcited charge accumulation and rates of interfacial charge transfer, to the best of our knowledge, is unprecedented. To quantify the rate of electron transfer per ZnO NP per second, energy transfer from ZnO to a bright organic fluorophore, toluidene blue, was measured spectroscopically. Although photodeposition of Au NPs at defect states in ZnO could not be directly confirmed in this study, it is likely that accumulated electronic charge migrated to these sites and reduced AuCl2−, based on patterns of Au NP deposition on ZnO NP edge sites that were observed in highresolution transmission electron microscopy (HRTEM) measurements. Finally, TEM measurements confirmed that patterns of Au NP deposition depended on the light irradiation technique and dissolved oxygen content and were highly sensitive to the solvent system. These experimental parameters determined the rate of electron accumulation and interfacial electron transfer in large pristine ZnO NPs and in the hybrids, which in turn determined photodeposited Au NP size distributions, Au NP clustering, Au NP morphologies, and Au NP number density per ZnO NP in the hybrids produced. The significant degree of control achieved over Au photodeposition provides an important means to modify plasmonic and electronic properties for various photocatalytic applications.

2. EXPERIMENTAL SECTION 2.1. Materials. Zinc acetate dihydrate [Zn(OAc)2·2H2O, 98%], gold(III) chloride trihydrate (HAuCl4·3H2O, 99.99%), oleylamine (cis1-amino-9-octadecene, 70%), 1-dodecanol [CH3(CH2)11OH, 98%], 1dodecanethiol [CH3(CH2)11SH, 98%], cyclohexane (C6H12, 99%), and acetonitrile (CH3CN, 99.8%) were purchased from Sigma− Aldrich. Butanol [CH 3 (CH 2 ) 3 OH, 99.4%] and tert-butanol [(CH3)3COH, 99%] were purchased from Ajax Finechem. Ethanol (CH3CH2OH, absolute), toluene (C6H5CH3, 99.5%), ethyl acetate (CH3COOC2H5, 99%), and potassium hydroxide (KOH, 85%) were purchased from chem-supply. Toluidine blue was purchased from Proscitech. Methanol (CH3OH, 99.5%) was purchased from Merck. 2.2. Synthesis of ZnO Nanoparticles. Prolate ZnO NPs capped with oleylamine were prepared by a solvothermal method as reported by Li et al.50 with some modifications. In a typical synthesis, a precursor solution containing 6 mmol of Zn(OAc)2·2H2O, 6 mL of dodecanol, and 4 mL of oleylamine was prepared. The reaction mixture was refluxed at 175 °C for 3 h and then allowed to cool to room temperature. Finally, the as-obtained products were repeatedly washed with ethanol and dried in air. Synthesis of ZnO NPs without capping ligand (uncapped) is reported elsewhere.51 Briefly, 25 mmol of Zn(OAc)2·2H2O and 3.5 mmol of KOH were mixed in a mixture of methanol (45 mL) and ultrapure water (5 mL). To obtain prolate NPs, the reaction mixture was refluxed at 85 °C for 48 h. The as-

3. RESULTS AND DISCUSSION Photolysis of the Au salt precursor, gold(III) chloride trihydrate, and effects of preirradiation of the Au salt by 370 nm diode light output on rates of Au NP photodeposition on ZnO were first examined (section 3.1). Next the effects of irradiation technique (either pulsed or continuous), capping ligand, and dissolved oxygen content on the photodeposition process were investigated (sections 3.2−3.4). The effect of ε on Au NP growth rate and rate of interfacial electron transfer from ZnO to TB+ were investigated (section 3.5). Finally, attachC

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. TEM images of ZnO−Au product after pulsed 20 s exposure of 370 nm light for ZnO NPs in ethanol with (a) 100%, (b) 50%, and (c) 0% photolyzed HAuCl4 in ethanol and ZnO NPs in toluene with (d) 100%, (e) 50%, and (f) 0% photolyzed HAuCl4 in ethanol.

prereactive species, as has been outlined by Eustis et al.54.55 and others.56,57 Solutions of HAuCl4 (0.1375 mM in ethanol) were prepared and 0.6 mL was added to 1.25 mL of different pure solvents examined in this photodeposition study before their exposure to 370 nm UV light. Again, there was no evidence for photogeneration of Au NPs observed in the UV−vis spectra of these solutions after 10 min of irradiation (Figure S2), in the absence of ZnO NPs. Thus, it was concluded that the direct photolysis pathway for conversion of Au3+ to Au NPs was neither fast nor efficient and so was unlikely to compete with heterogeneous photodeposition of Au onto ZnO at the time scales over which Au NP photodeposition was investigated. 3.1.2. Slow Room-Light Deposition onto ZnO. Although ZnO requires UV light for efficient photogeneration of electrons, it may act as a photocatalyst under ambient light illumination. To test this possibility, 0.1375 mM solutions of HAuCl4 (in ethanol) were irradiated with 370 nm light for 6 min and then added to ZnO colloids suspended in several different solvents (ethanol, toluene, and others). Reduction of the Au precursor was observed in TEM samples with all solvent mixtures after 10 min, although for the ZnO suspensions in ethanol, only minimal Au NP deposition occurred under room light and ambient conditions (Figure S3a). This confirmed that preirradiation could be used to generate excited-state AuCl4−* of the auric complex, which, following homolysis of the Au−Cl bond, transformed into AuCl3− and the metastable disproportionation product AuCl2− that could readily reduce at the ZnO surface, thus activating the gold salt prior to heterogeneous deposition onto ZnO NPs. When the preirradiated HAuCl4 solutions were combined with ZnO suspensions in toluene, it was observed that a distinctive Au NP plasmon peak appeared in the UV−vis spectrum after 10 min of exposure to room light at room temperature (Figure S4). Au NP deposition was clearly faster in a toluene/ethanol mixture compared to the same reaction in ethanol solvent under room light at room temperature. TEM analysis of the products revealed that Au NP sizes within these samples covered a broad range between 2

ment of the photodeposited Au NPs to ZnO was examined in close detail by HRTEM (section 3.6). 3.1. Effects of Preirradiation of HAuCl4 Solution. 3.1.1. Control Experiments. It has been noted previously that photolysis of Au salts is frequently overlooked during Au NP preparation.52 Photolysis of AuCl4− generates AuCl3− and a highly reactive free radical chlorine species as products.53 The AuCl3− species rapidly disproportionates to re-form AuCl4− and generate AuCl2−, a metastable intermediate that can undergo further slow disproportionation to form Au0.54−57 In addition, the reactive Cl free radical can abstract hydrogen from alcohols or other organics to generate powerful reducing free radical species that can produce Au NPs.52 Depending upon the efficiency of HAuCl4 photolysis product free radical side reactions, these could compete with direct photoreduction on ZnO. The possibility for competitive side reactions interfering with direct photodeposition was therefore first tested under room lights at room temperature by monitoring the UV−vis spectrum. Under these conditions, photoreduction of HAuCl4 to form Au NPs was not detected in the UV−vis spectrum at ∼530 nm, where the Au NP plasmon absorption peak might be expected to appear. Thus, consistent with previous studies, there was no evidence to suggest that ethanolic solutions of HAuCl4 decomposed under visible light to form Au NPs on the time scales over which Au NP photoreductions on ZnO were investigated here.52 Similarly, even in the presence of ZnO photocatalyst, in ethanol and under the same ambient and room light conditions, heterogeneous photoreduction of HAuCl4 on ZnO could not be discerned in the UV−vis spectrum, and no Au NP plasmon peak was detected after 10 min. Next, the stability of UV-preirradiated HAuCl4 was tested when it was added to different neat solvents. Since the wavelength range of the 370 nm UV diode light source used in these experiments overlaps with the AuCl4− ligand-to-metal charge transfer (LMCT) absorption band at 320 nm (Figure S1), the precursor salt could be converted to AuCl2− and other D

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

number (particle size ranged from 3 to 10 nm) and many of the ZnO NPs did not form hybrids, remaining bare of Au NPs. 3.2. Effect of Irradiation Technique. The irradiation technique, pulsed or continuous light exposure, also appeared to play an important role in determining the final form of ZnO−Au hybrid product (in combination with the type of solvent used). Figure 2a,b shows TEM images of the ZnO−Au

and 10 nm. Many were also obviously attached to the ZnO NPs (Figure S3b). In toluene solvent, very slow deposition of Au NPs was observed even under dark conditions. From these experiments, it was clear that UV light preirradiation of the HAuCl4, addition to ZnO converts prereactive AuClx into Au NPs, but only at a slow rate. 3.1.3. Rapid UV Photodeposition onto ZnO and Effect of Varying Preirradiation Time. Photodeposition onto ZnO initiated by 370 nm irradiation was expected to be more rapid. To confirm this hypothesis, preirradiated ethanolic HAuCl4 solutions were mixed with suspensions of ZnO NPs in ethanol and then given a short 20 s pulsed exposure with 370 nm light. The Au NP deposition process was monitored by observing the appearance and growth of the Au NP plasmon peak in the UV−vis spectrum at ∼530 nm during the light pulse. TEM images confirmed that this photodeposition procedure generated Au NPs of approximately 14 nm diameter on ZnO and formed hybrid particles (Figure 1a). When the same experiment was performed in toluene solvent, TEM measurements confirmed that the ZnO particles were covered in Au NPs with markedly smaller sizes (∼6 nm) compared to ethanol (Figure 1d). The pulsed UV irradiation also nucleated and photodeposited a far greater density of Au NPs on ZnO in toluene compared with products obtained in ethanol. In each case, photodeposition with UV light was much faster than deposition under room light. More importantly, the UV photodeposition process (which occurs within 10 s of UV light as shown in section 3.2) was much faster than the preirradiation time of HAuCl4 (∼6 min). This suggests that if the preirradiation time were decreased, the photodeposition rate (and hence the particle size and coverage) would decrease. Preirradiation time could therefore be one parameter that can be used to control Au NP size and number density, in a process that is also solvent-dependent. The effect of different levels of preirradiation of ethanolic HAuCl4, prior to addition to capped ZnO was studied in detail in both ethanol and toluene. By use of the 370 nm light source, the pulsed 20 s photodeposition process was compared for the cases of (i) complete (100%) preirradiation photolysis of HAuCl4, (ii) 50% photolysis, and (iii) no photolysis of the Au precursor prior to its addition to ZnO colloid. By monitoring the AuCl4− precursor metal-to-ligand charge transfer (MLCT) absorption band at 320 nm, it could be established when 50% and 100% photolysis occurred (Figure S1). The first set of photodeposition experiments on capped ZnO in ethanol demonstrated that preirradiation had a significant effect on Au NP size distribution in the final product and on the number of Au NPs deposited per ZnO NP. This can be seen qualitatively by observing the TEM images of Figure 1a−c, where 100% Au salt preirradiation photolysis led to multiple Au NPs nucleating on each ZnO NP in Figure 1a compared to the isolated Au NP photodeposition on ZnO that occurred for 50% and 0% Au salt preirradiation photolysis in Figure 1b,c. In a similar trend, the number of Au NPs photodeposited per ZnO in toluene was also dependent on Au precursor preirradiation time, as can be seen by examination of Figure 1d−f. After the reaction was performed with a 50% photolyzed AuCl4− solution, larger Au NPs were observed to photodeposit on the capped ZnO NPs (Figure 1e). In this case the Au NPs were approximately 8 nm in diameter and there were fewer Au NPs deposited. Figure 1f is a TEM image of Au photodeposited on ZnO without the preirradiation step in toluene solvent. Clearly Au NPs photodeposited, but they were far fewer in

Figure 2. (a, b) ZnO−Au hybrid product obtained after continuous irradiation of ZnO in (a) ethanol and (b) toluene. (c) Second derivative of extinction (with respect to time at the plasmon peak wavelength) for ethanol and toluene in a pulsed irradiation experiment. For more details, see Supporting Information. Note: the slight apparent delay in UV pulse time likely results from the large time steps used.

hybrid product obtained with oleylamine-capped ZnO in ethanol and toluene, after 20 s of continuous exposure to 370 nm diode irradiation. This can be compared with hybrid product produced by 4 × 5 s pulses of 370 nm diode irradiation in Figure 1a,d in ethanol and toluene solvents, respectively. In ethanol, continuous irradiation produced Au nanoclusters rather than spherical NPs at the ZnO surface. In toluene, comparison of the TEM results in Figures 1d and 2b shows that there is no apparent difference in the products obtained after pulsed and continuous irradiation. The likely reason for this is that photochemical reduction in toluene completes in less than 5 s, as can be seen by inspection of the second derivative of extinction illustrated in Figure 2c, whereas in ethanol the reaction completes after approximately 8 s of exposure to UV light. Therefore, in ethanol solvent during continuous irradiation, photoexcited electron accumulation occurred, generating Au clusters, whereas during pulsed irradiation, electron discharge occurred between pulses to generate spherical Au NPs. Extinction spectra for pulsed and continuous irradiation exposures in ethanol are shown in Figure 3a,b. The wavelength corresponding to the maximum of the collective Au plasmon peak is not wholly related to the size of photodeposited Au NPs, because the change in extinction for spectra of these hybrids is not a simple linear combination of ZnO and Au E

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

instead. These larger, more complex Au nanostructures had a plasmon extinction peak at ∼610 nm, which gave the solution a blue color (see inset of Figure 3b). This type of UV−vis extinction (Figure 3b) has frequently been observed in Au NP clusters.58 Unlike the ethanol solvent results, extinction spectra in toluene for pulsed (Figure 3c) and continuous irradiation exposures were very similar. The newly formed hybrid NPs generated a Au plasmon peak extinction maximum at ∼550 nm. As indicated by the results in Figure 2c, in toluene, photodeposition appeared to complete within the first 5 s of the light exposure pulse, so no significant change in plasmon peak developed after 5 s. 3.3. Effect of ZnO Capping Ligand. As discussed in the Experimental Section, Au photodeposition was investigated on both uncapped and oleylamine-capped ZnO samples. Comparison of Figures 1a and 4a, for the case of photoreduction of

Figure 4. Effect of capping ligand: ZnO−Au hybrid product formed after pulsed irradiation of a mixture of preirradiated HAuCl4 and (a) uncapped ZnO dispersed in ethanol or (b) DDT-capped ZnO dispersed in toluene.

preirradiated HAuCl4 on these two different ZnO sample types in ethanol solvent, shows that apart from minor differences in the original ZnO NPs morphology and size distributions produced during preparation, the Au photodeposited products appeared indistinguishable. After 20 s pulsed irradiation with 370 nm UV light, ∼14 nm Au NPs were observed to nucleate and grow at the same types of sites on these NPs, at crystal facet edges and tips. In all the solvent systems studied and irradiation conditions explored, capping with the labile and soft oleylamine ligand was not observed to change the photodeposition when these two ZnO NP types were compared (Figure S5). Photodeposition onto DDT-capped ZnO was also attempted and explored (Figure 4b). DDT-capped ZnO was readily dispersed in toluene solvent, and a number of irradiation wavelengths, irradiation times, and solvent systems were investigated. Consistently, Au did not photodeposit to produce hybrid NPs. Clearly the DDT shell surrounding these ZnO NPs acted as a barrier for interfacial electron transfer and subsequent photoreduction of AuCl2−. DDT has previously been observed to form a compact, disordered monolayer that passivates surface defect sites of ZnO nanocrystals.59,60 3.4. Effect of Dissolved Oxygen Content. 3.4.1. Au Nanocluster Formation from Electron Accumulation. At standard temperature and pressure (STP), both ethanol and toluene naturally contain similar concentrations of dissolved oxygen (2.10 and 1.83 mM, respectively)61,62 that can act as an electron scavenger of photoexcited e−CB from ZnO colloidal particles.47 It was reported previously that the rate of reaction of oxygen with e−CB is relatively slow compared to electron

Figure 3. Evolution of extinction spectra of Au nanostructures depositing on ZnO NPs with time in (a) pulsed irradiation in ethanol, (b) continuous irradiation in ethanol, and (c) pulsed irradiation in toluene.

extinction spectra.2 Nevertheless, clear trends in Au NP growth can be interpreted. During the pulsed irradiation in Figure 3a, a broad plasmon peak appeared within the first 5 s irradiation pulse. This broad peak can be a signature of the formation of tiny Au NP clusters on the ZnO NPs. After this initial hybrid formation, the results displayed in Figure 3a indicate that, during the UV-off period, a significant blue shift in the Au plasmon peak extinction wavelength occurs toward ∼530 nm. This could be attributed to ripening of small Au clusters into larger, spherical Au NPs on the ZnO NP surface. A similar blue shift in the plasmon extinction peak during Au NP growth was not observed with continuous irradiation in ethanol solvent. Under these irradiation conditions, large Au clusters formed F

DOI: 10.1021/acsami.6b03128 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces generation, and even slower in the case of larger particles.46 Under UV irradiation, assisted by hole scavenging by the ethanol present in these solutions, excited electrons can transfer to Au complex species such as AuCl2− ions and to dissolved oxygen in the solvent. The dissolved oxygen present in the solvent scavenges electrons and so minimizes accumulation of e−CB in ZnO, leading to formation of spherical Au NPs, as was observed in Figure 1a,d under pulsed UV irradiation. Removal of the dissolved oxygen by bubbling argon through the solution for a minimum of 20 min led to a completely different pattern of Au photodeposition on ZnO. After elimination of dissolved oxygen by this method, photoexcited electrons could accumulate in the conduction band of ZnO NPs during exposure to the 370 nm photodiode output. These accumulated electrons can be expected to create a substantial potential gradient at the semiconductor−solution interface and attract a high proportion of AuCl2− ions, which reduce to form large Au clusters. Au cluster formation of this type can be seen in Figure 5a,b with ZnO NPs dispersed in either ethanol or toluene

was performed under ambient conditions (compare inset of Figure 5c to inset of Figure 3c). 3.4.2. Electron Accumulation Studies That Use Electron Paramagnetic Resonance. EPR spectroscopy is a useful technique to demonstrate accumulation of excited electrons in semiconductor NPs. Schimpf et al.47 reported the appearance of an EPR signal at g = 1.96 for quantum-sized ZnO (